Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add instructions/edwardxckf/humanizer-ms/claude-mdgit clone --depth 1 https://github.com/edwardxckf/humanizer-msWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/edwardxckf/humanizer-ms/claude-md)<a href="https://agentmods.dev/instructions/edwardxckf/humanizer-ms/claude-md"><img src="https://agentmods.dev/badge/instructions/edwardxckf/humanizer-ms/claude-md.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.01138 | $0.01138 |
| Opus 5 | $0.00569 | $0.00569 |
| Sonnet 5 | $0.00228 | $0.00228 |
| Haiku 4.5 | $0.00114 | $0.00114 |
Grade A, and why
humanizer-ms CLAUDE.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 73 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
What this repo is
A content repo, not a code repo. There is no build, no test suite, no lint, no dependencies. The deliverable is an agent skill that removes AI writing patterns from Malay (Bahasa Melayu) text, adapted from blader/humanizer (English) and op7418/humanizer-zh (Chinese).
Architecture: two payload files, everything else is packaging
Everything under skills/humanizer-ms/ is the payload. Every other file exists only to get that
folder in front of some agent harness.
| Payload file | Role |
|---|---|
SKILL.md |
The patterns. Frontmatter description is what triggers the skill; body holds 34 numbered patterns with before/after examples in Malay, detection guidance (false positives, signs of human writing), invocation modes, and the draft → audit → final loop |
kata-ai.txt |
One watched phrase per line, consumed by grep -oiFf. A smell counter, not a ban list. Deliberately short: only phrases that are wrong in every register, so the grep does not flag correct formal Malay |
The two travel together: AGENTS.md and the READMEs shell out to kata-ai.txt by relative path.
Packaging fan-out — edit one, check the others
The same skill is registered six ways. A change to the skill's name, description, or path has to be mirrored:
.claude-plugin/plugin.json+.claude-plugin/marketplace.json— Claude Code.codex-plugin/plugin.json— Codex (also carriesinterface.longDescriptionanddefaultPrompt).agents/plugins/marketplace.json— Codex/Antigravity; hardcodes the upstream clone URL, so a fork that doesn't change the owner here will install upstream instead of itselfskills/humanizer-ms/agents/openai.yaml— OpenAI-harness interface blockgemini-extension.json→GEMINI.md→@./skills/humanizer-ms/SKILL.md(import chain)AGENTS.md— plain rules-file route; duplicates the grep commandplugin.jsonat repo root — minimal name/description
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 5d ago First seen · 73 lines · 1,138 tokens per session scan A 4db244fe0da3
humanizer-ms CLAUDE.md is an instructions file published in the GitHub repository edwardxckf/humanizer-ms (2 stars, last pushed 18d ago), licensed MIT. It adds 1,138 tokens to every session, about $0.0057 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
avoid-ai-writing-multilingual CLAUDE.md
Claude Code instructions for jurigis/avoid-ai-writing-multilingual, covering avoid-ai-writing — multilingual skills, project structure, how to create a new language skill, step 1: research sources (web search — not from memory) and step 2: identify language-specific categories.
im-not-ai GEMINI.md
Instructions for epoko77-ai/im-not-ai, covering humanize kr — ai 한글 티 제거 (gemini cli extension), 개요, 커스텀 명령, 철칙 (위반 시 즉시 롤백) and do-not (탐지·윤문 모두 제외).
humanizer-zh-tw AGENTS.md
Instructions for nagameTW/humanizer-zh-tw, covering agents.md, 這個 repo 是什麼, 關鍵檔案, 維護契約（改東西前先讀） and 編輯 skill.md.
editorial-skill-for-natural-russian-business-writing AGENTS.md
Instructions for ValentinAvramko/editorial-skill-for-natural-russian-business-writing, covering инструкции для агента, язык и документация, границы документов, источник правды and git.
shuorenhua-zh-tw AGENTS.md
Instructions for tentenco/shuorenhua-zh-tw, covering 倉庫說明, 語言, 風格, 布局 and 協作.
shuorenhua-zh-tw CLAUDE.md
Instructions for tentenco/shuorenhua-zh-tw, a project described as: 繁體中文（臺灣）AI 味清理 Skill：去 AI 腔、清中港用語、修簡轉繁錯字、教育部標點、七種出版語體。AI humanizer skill for Traditional Chinese (Taiwan) — Claude Code / Codex / Cursor / ChatGPT.